Skip to content

help for beginner

5 messages · Nguyen Chanh Tu, Albyn Jones, Fikret ER +2 more

Original
Albyn Jones
# 
Have you looked at the teaching resources on the CRAN website?
If not, take a look at the "Contributed" link under Documentation.
There are also numerous published texts that you might find helpful,
eg as sources of examples.  Many of the help files for specific 
functions
have examples, and R has numerous datasets, both in the base package 
and in
other packages.

albyn
On 2012-11-18 18:09, Nguyen Chanh Tu wrote:
Fikret ER
# 
Hi
There are few things that you can do,


1. Install the package "TeachingDemos" by Greg Snow,

2. Visit the Rwiki page, http://rwiki.sciviews.org/doku.php

3. May visit the page http://www.cyclismo.org/tutorial/R/probability.html

4. I also find R Journal very useful, http://journal.r-project.org

5. For Introductory Statistics, the book "Introductory Statistics with R"
by Peter Dalgaard is excellent choice,

6. May read the slides of S. Weisberg,
http://www.r-project.org/user-2006/Slides/Fox+Weisberg.pdf

7. Visit the page here,
http://www.r-bloggers.com/teaching-with-r-the-tools/


I hope these will help,

Fikret ER
Anadolu University
Eskisehir
Turkey

Fer at anadolu.edu.tr
On 11/19/12 6:23 AM, "Albyn Jones" <jones at reed.edu> wrote:

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Randall Pruim
    

    

      
      #
    

  


  

      
I haven't taught from that book in a while, but here are some tips:

1) If the 7th edition is still current, grab the Devore7 package which contains many of the data sets (sometimes awkwardly formatted and poorly labeled, but at least they are there.)

2) Choose a graphics system.  I recommend lattice, but I really recommend that you don't mix and match.  One advantage of lattice is that it teaches the formula interface you will want to use for linear models, so students will be a step ahead when you get there.

3) If you go with lattice, I suggest you also give the mosaic package a try.  It adds other functionality via a formula interface (e.g. mean(age ~ sex, data=foo) gives mean ages broken down by sex) and simplifies some important tasks, like adding a linear model fit to a plot or evaluating the fit at specific values of the explanatory variables.

4) You might also take a look at the minimalR vignette in the mosaic package.  It was designed for a slightly different course, but tries to give a small but sufficient set of commands for teaching introductory statistics using lattice and mosaic.  I'd encourage you to find or create something like this for your students.  (The target audience for the current vignette is really faculty.)

5) Less volume, more creativity.

Try to keep the number of different R commands you teach to a minimum.  Choose them wisely and choose tools that play well together.  That makes it easier for the students to master the skills they really need.  R often provides several ways to skin a cat, but your students only need one way.

Have fun.

---rjp

        
On Nov 18, 2012, at 9:09 PM, Nguyen Chanh Tu wrote:

        

            

      
      
      
    

        

      
      
    

  
  


  


      

    

    
9 days later

    

      
      

        


  

        

      Robert Baer
    

    

      
      #
    

  


  

      
On 11/18/2012 8:09 PM, Nguyen Chanh Tu wrote:

            

      
      
      
    

        

      
      
    

        
I have used R to teach biomedical science students for several years, 
and although not without some frustrations, students catch on fairly 
quickly.  I don't have experience with engineering students, so their 
needs may differ.

My approach is to spend a couple of class periods at the beginning 
making sure that the students can enter a data set in one of several 
ways, for example, using read.table(), scan(), and using the c() with 
assignment.  We spend time with the concept of a vector and a dataframe 
and the difference in how a one column dataframe and a vector are 
printed at the command prompt.  We talk about how to move back and forth 
between vectors and dataframes.  We talk about how to change one value 
in a mistyped vector value and how to change one value in a dataframe 
column and row.  I do this despite the fact that virtually all the REAL 
datasets we use are available electronically because it puts the student 
in control of sometimes "assumed" ideas of R users.  I also provide 
students with a two-sided cheat-sheet of the most common commands that I 
think we will use.  I invite them to help me keep it current.  I 
regularly devote class time to doing live demonstrations on "sample 
problems" similar to those I want them to be able to use R for.  My 
students have computers and can work along when we do this.  Time spent 
in these endeavors seems far more important than it would using an Excel 
interface.

We also talk about the difference between the so-called "long" and 
"wide" formats of dataframes and introduce them to working with factors 
and the very strange (to the newcomer) as.numeric(as.character(factvar)) 
construction.  I am frequently surprised at how much more difficult 
these ideas of data layout are than the statistics or R commands 
themselves are for the students. Yet, this skill-set is highly 
transferable across programs they might encounter in the future.  I 
unashamedly look at the help files for commands we are using (I do this 
during class) even when we are not doing anything too tricky.  This role 
models the most basic R support for the students, right from the start.

Someone already mentioned R Commander which I only introduce to the 
students when we get to ANOVA and linear regression.  By then they have 
a couple of weeks getting used to the feel of a command line interface.  
Most of my students rarely look back from here, but for engineering 
students R is often a useful free substitute for Matlab capabilities, 
and so the more fundamental skills you can convey to the students, the 
better off they will be (in my opinion).

Another very nice GUI interface for R is Deducer, particularly when used 
in conjunction with JGR.  If all you want to do is to teach "basic 
probability and statistics", this might be another way to go.

Rob Baer